Space As a Key Word

1

SPACE AS A KEY WORD

David Harvey

Paper for Marx and Philosophy Conference, 29 May 2004, Institute of Education, London

DRAFT VERSION ONLY

The author welcomes comments and feedback on this draft please email to

If Raymond Williams were contemplating the entries for his celebrated text on Keywords today, he would surely have included the word “space.” He may well have included it in that short list of concepts, such as “culture” and “nature”, to be listed as “one of the most complicated words in our language.” How, then, can the range of meanings that attach to the word “space” be expanded upon and clarified without losing ourselves in some labyrinth (itself an interesting spatial metaphor) of complications?

Space is, of course, one of those words that frequently elicits modification. The complications perhaps arise more out of the modifications (which all too frequently get omitted in the telling or the writing) rather than out of any inherent complexity of the notion of space itself. When, for example, we write of “material”, “metaphorical”, “liminal”, “personal”, “social” or “psychic” space (just to take a few examples) we thereby indicate a considerable diversity of contexts which so inflect matters as to seem to render the meaning of space itself entirely contingent upon the context. Similarly, when we designate its range of applications in terms such as spaces of fear, of play, of cosmology, of dreams, of anger, of particle physics, of capital, of geopolitical tension, of hope, of memory, or of ecological interaction (again, just to indicate a few of a seemingly infinite range of potential sites of deployment of the term) then we seem to be saying that the arena of application defines something so special about the meaning of space as to render any general consideration of its properties a hopeless task.

Important and influential though all these modifiers are, some more general exploration of the meaning of the term seems necessary. Only then will we be able to disperse some of the fog of mis-communication that seems to bedevil use of the word.

It is, however, difficult to decide where to begin. That determination is not innocent since how enquiry is stituated (to deploy yet another spatial metaphor) defines a particular perspective that highlights some matters while occluding others. And while it may be correct in the long run to accord a certain privilege to the realm of, say, philosophical reflection, as an arena of thought whose mission is primarily to rise above the various and divergent fields of human practices and partial knowledges to tell us something more definitive about the categories to which we may appeal, I have formed the impression that there is sufficient dissension and confusion amongst the philosophers as to make that anything but an unproblematic starting point. Furthermore, since I am by no means qualified to contemplate the concept of space from within the philosophical tradition, it seems best to begin where I can best begin and then see what others make of it. I therefore begin from the standpoint of the geographer, not because this is a privileged site that somehow has a proprietary right (as some geographers sometimes seem to claim) over the use of spatial concepts, but because that is where I happen to be and it is in this arena where I have had to wrestle most directly with the complexity of what the word space might be all about. I have, of course, frequently sought assistance and inspiration from others operating elsewhere in the complex division of academic and practical labor. But my aim in appropriating the ideas of others has never been some grand synthesis of meanings. I have sought, rather, meanings that work, as satisfactorily as possible, for the theoretical and practical topics of primary concern to me.

I began reflecting upon this problem many years ago. In Social Justice and the City, published in 1973, I argued that it was crucial to reflect on the nature of space if we were to understand urban processes under capitalism. Drawing upon ideas previously culled from a study of the philosophy of science I identified a tripartite division in the way space could be understood:

“If we regard space as absolute it becomes a “thing in itself” with an existence independent of matter. It then possesses a structure which we can use to pigeon-hole or individuate phenomena. The view of relative space proposes that it be understood as a relationship between objects which exists only because objects exist and relate to each other. There is another sense in which space can be viewed as relative and I choose to call this relational space - space regarded in the manner of Leibniz, as being contained in objects in the sense that an object can be said to exist only insofar as it contains and represents within itself relationships to other objects.”

I think this tripartite division is well-worth sustaining not so much as an abstraction but as a mix of means to understand events occurring around us and to formulate ways of thinking and theorizing about geographical phenomena and processes. The arguments I have had with architects and with sociologists and others over the years often boil down, I have found, to arguments predicated on looking at the nature of space in these different ways.

So let me begin with a brief elaboration on what each of these categories might entail. Absolute space is fixed and we record or plan events within its frame. This is the space of Newton and Descartes and it is usually represented as a pre-existing and immoveable grid amenable to standardized measurement and open to calculation. Geometrically it is the space of Euclid and therefore the space of all manner of cadastral mapping and engineering practices. It is a primary space of individuation - res extensa as Descartes put it - and this applies to all discrete and bounded phenomena including you and me as individual persons. Socially this is the space of private property and other bounded territorial designations (such as states, administrative units, city plans and urban grids). When Descartes’ engineer looked upon the world with a sense of mastery, it was a world of absolute space (and time) from which all uncertainties and ambiguities could in principle be banished and in which human calculation could uninhibitedly flourish.

The relative notion of space is mainly associated with name of Einstein and the non-Euclidean geometries that began to be constructed most systematically in the 19th century. Space is relative in the double sense: that there are multiple geometries from which to choose and that the spatial frame depends crucially upon what it is that is being relatavized and by whom. When Gauss first established the rules of a non-Euclidean spherical geometry to deal with the problems of surveying accurately upon the curved surface of the earth, he also affirmed Euler’s assertion that a perfectly scaled map of any portion of the earth’s surface is impossible (incidentally validating Borges’ later fantasy of the perfect map that in order to guarantee perfection has to be exactly the same scale as the territory being mapped). Einstein took the argument further by pointing out that all forms of measurement depended upon the frame of reference of the observer. The idea of simultaneity in the physical universe, he taught us, has to be abandoned. It is impossible to understand space independent of time under this formulation and this mandates an important shift of language from space and time to space-time or spatiotemporality. It was, of course, Einstein’s achievement to come up with exact means to examine such phenomena as the curvature of space when examining temporal processes operating at the speed of light. But in Einstein’s schema time remains fixed while it is space that bends according to certain observable rules (much in the same way as Gauss devised spherical geometry as an accurate means to survey through triangulation on the earth’s curved surface).

At the more mundane level of geographical work, we know that the space of transportation relations looks and is very different from the space of private property relationships. The uniqueness of location and individuation defined by bounded territories in absolute space gives way to a multiplicity of locations that are equi-distant from, say, some central city location. We can create completely different maps of relative locations by differentiating between distances measured in terms of cost, time, modal split (car, bicycle or skateboard) and even disrupt spatial continuities by looking at networks, topological relations (the optimal route for the postman delivering mail), and the like. We know, given the differential frictions of distance encountered on the earth’s surface, that the shortest distance (measured in terms of time, cost, energy expended) between two points is not necessarily given by the way the legendary crow flies over physical distance. Furthermore the standpoint of the observer plays a critical role. The typical New Yorker’s view of the world, as the famous Steinberg cartoon suggests, fades very fast as one thinks about the lands to the west of the Hudson River or east of Long Island.

All of this relativization, it is important to note, does not necessarily reduce or eliminate the capacity for calculability or control, but it does indicate that special rules and laws are required for the particular phenomena and processes under consideration. Difficulties do arise, however, as we seek to integrate understandings from different fields into some more unified endeavor. The spatiotemporality required to accurately capture energy flows through ecological systems, for example, may not be compatible with that of financial flows through global markets. Understanding the spatio-temporal rhythms of capital accumulation requires a quite different framework to that required to understand global climate change. Such disjunctions, though extremely difficult to work across, are not necessarily a disadvantage provided we recognize them for what they are. Comparisons between different spatio-temporal frameworks can illuminate problems of political choice (do we favor the spatiotemporality of financial flows or that of the ecological processes they typically disrupt, for example).

The relational concept of space is most often associated with the name of Leibniz who, in a famous series of letters to Clarke (effectively a stand-in for Newton) objected vociferously to the absolute view of space and time so central to Newton’s theories. The relational view of space holds there is no such thing as space outside of the processes that define it. Processes do not occur in space but define their own spatial frame. The concept of space is embedded in or internal to process. This very formulation implies that, as in the case of relative space, it is impossible to disentangle space from time. We must therefore focus on the relationality of space-time rather than of space in isolation. The relational notion of space-time implies the idea of internal relations; external influences get internalized in specific processes or things through time (much as my mind absorbs all manner of external information and stimuli to yield strange patterns of thought including dreams and fantasies as well as attempts at rational calculation). An event or a thing at a point in space cannot be understood by appeal to what exists only at that point. It depends upon everything else going on around it (although in practice usually within only a certain range of influence). A wide variety of disparate influences swirling over space in the past, present and future concentrate and congeal at a certain point to define the nature of that point. Identity, in this argument, means something quite different from the sense we have of it from absolute space. Thus do we arrive at an extended version of Leibniz’s concept of the monad.

Measurement becomes more and more problematic the closer we move towards a world of relational space-time. But why would it be presumed that space-time only exists if it is measurable and quantifiable in certain traditional ways? This leads to some interesting reflections on the failure (perhaps better construed as limitations) of positivism and empiricism to evolve adequate understandings of spatio-temporal concepts beyond those that can be measured. In a way, relational conceptions of space-time bring us to the point where mathematics, poetry, and music converge. And that is anathema to those of a positivist or crudely materialist bent. But Leibniz’s return to popularity and significance not only as the guru of cyberspace but also as a foundational thinker in relationship to more dialectical approaches to mind-brain issues and quantum theoretical formulations signals some sort of urge to go beyond absolute and relative concepts and their more easily measurable qualities. But the relational terrain is an extremely challenging and difficult terrain upon which to work. There are, however, many thinkers who, over the years, have applied their talents to reflecting upon the possibilities of relational thinking. Deleuze, for one, has made much of these ideas both in his reflections on Leibniz (with reflections on baroque architecture and the mathematics of the fold in Leibniz’s work) as well as of Spinoza. Here he is, for example, writing on the latter:

“The important thing is to understand life, each living individuality, not as a form, but as a complex relation between different velocities, between decelerations and accelerations of particles. A composition of speeds and slownesses on a plane of immanence.”

But why would I, as a mundane working geographer, find the relational mode of approaching space-time useful? The answer is quite simply that there are certain topics, such as the political role of collective memories in urban processes, that can only be approached this way. I cannot box political and collective memories in some absolute space (clearly situate them on a grid or a map) nor can I understand their circulation according to the rules, however sophisticated, of relative space-time. If I ask the question: what does Tiananmen Square or “Ground Zero” mean, then the only way I can seek an answer is to think in relational terms.

So is space absolute, relative or relational? I simply don’t know whether there is an ontological answer to that question. Even if there were I do not have the intellectual means to make or even evaluate that determination. In my own work I think of it as being all three. This was the conclusion I reached thirty years ago and I have found no particular reason (nor heard any arguments) to make me change my mind. This is what I wrote in 1973:

“space is neither absolute, relative or relational in itself, but it can become one or all simultaneously depending on the circumstances. The problem of the proper conceptualization of space is resolved through human practice with respect to it. In other words, there are no philosophical answers to philosophical questions that arise over the nature of space - the answers lie in human practice. The question “what is space?” is therefore replaced by the question “how is it that different human practices create and make use of different conceptualizations of space?” The property relationship, for example, creates absolute spaces within which monopoly control can operate. The movement of people, goods, services, and information takes place in a relative space because it takes money, time, energy, and the like to overcome the friction of distance. Parcels of land also capture benefits because they contain relationships with other parcels….in the form of rent relational space comes into its own as an important aspect of human social practice.”

Are there rules for deciding when and where one spatial frame is preferable to another? Or is the choice arbitrary, subject to the whims of human practice? The decision to use one or other conception certainly depends on the nature of the phenomena under investigation. The absolute conception may be perfectly adequate for issues of property boundaries and border determinations but it helps me not a whit with the question of what is Tiananmen Square or Ground Zero. I therefore find it helpful - if only as an internal check - to sketch in justifications for the choice of an absolute, relative, or relational frame of reference. Furthermore, I often find myself presuming in my practices that there is some hierarchy at work among them in the sense that relational space can embrace the relative and the absolute, relative space can embrace the absolute, but absolute space is just absolute and that is that. But I would not confidently advance this view as a working principle let alone try to defend it theoretically. I find it far more interesting in principle to keep the three concepts in dialectical tension with each other and to think the interplay among them. Ground Zero is an absolute space at the same time as it is relative and relational in space-time. When I look at a house, for example, I recognize it as a physical and legal entity that situates it in absolute space. I also recognize its position in relative space given its location with respect to places of employment, recreation, services and the flows of people, electricity, water, and money that sustain it as a living habitat. But then I also understand its relationality to global property markets, changing interest rates, climatic change, the sense of what is or is not a historic building, and its significance as a place of personal and collective memories, sentimental attachments, and the like. What happens to the house over time can only be fully understood, I argue, by working through effects constituted through the three forms of spatio-temporality simultaneously. While this is hard to do in any easy empiricist or positivist sense, the insights that come from such a dialectical approach are as exciting and innovative as they are often stunning.